Methods: Between 1993 and 1994 16,780 German infants were recruited in a large pertussis vaccine efficacy study. Of these 74.6 % initially were vaccinated with acellular pertussis vaccine (Biken DTaP), 10.9 % with whole-cell pertussis vaccine and 14.5% were not vaccinated against B. pertussis. Other acellular pertussis vaccines were used after licensure for catch-up or booster vaccination. Prospective surveillance for B.pertussis and B.parapertussis in the initial cohort was reinitiated in July 1997 in a case-control study investigating the long-term efficacy of pertussis vaccines. All children presenting with cough of >= 7 days duration were investigated by culture and PCR, and in addition by single serum sample serology when cough lasted >= 21 days. Parents were asked to document specific clinical symptoms in cough diaries until at least 42 days after start of coughing Results: A total of 86 cases were identified by culture, PCR or serology. Of these 45 (52%) were caused by B.pertussis and 41 (48%) by B.parapertussis. 76% of the children with B.pertussis infection and 93% of the children with B.parapertussis infection had received >= 3 doses of a pertussis vaccine( either aP or wcP). In contrast to 72% of the unvaccinated children, only 37% of the vaccinated children with B. pertussis infection fulfilled typical clinical criteria for pertussis disease. Duration of cough, paroxysms and vomiting were not significantly different between B.pertussis and B. parapertussis cases, whereas whooping occurred significantly more often in B.pertussis cases. Conclusion: In a highly vaccinated population B.pertussis disease often presents as a milder cough disease, and cannot be distinguished on a clinical basis alone from cough diseases caused by other microorganisms, including B.parapertussis.